Abstract
In order to achieve high precision Question Answering Systems or Information Retrieval Systems, the incorporation of Natural Language Processing techniques are needed. For this reason, in this paper a method that can be integrated in these kinds of systems, is presented. The aim of this method, based on maximum entropy conditional probability models, is semantic role labelling. The method, named SemRol, consists of three modules. First, the sense of the verb is disambiguated. Then, the argument boundaries of the verb are determined. Finally, the semantic roles that fill these arguments are obtained.
Cite
CITATION STYLE
Moreda, P., Palomar, M., & Suárez, A. (2004). SemRol: Recognition of semantic roles. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 3230, pp. 328–339). Springer Verlag. https://doi.org/10.1007/978-3-540-30228-5_29
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
